Applicability Analysis of Different Runoff Schemes of the Gulang River Basin in Arid Region of Gansu Province, China

¹ PIESAT International Information Technology Limited, Beijing, 100195, China
² Tianjin Normal University, Tianjin, 300387, China

^* Corresponding author: ezhao_cn@163.com

Abstract

[Background] Mountain torrent disasters are one of the deadliest weather-related natural disasters in the world, often causing extremely serious economic and property losses and casualties; in recent years, due to heavy rainfall and complex geological and geomorphological conditions, mountain torrent disasters have occurred frequently in Gansu Province. An effective rainfall-runoff model is the key to prevent them. There are still many problems about how to establish suitable hydrological models in the arid basin of China. [Methods] Therefore, the Gulang River Basin in Gansu Province was selected as the research area to construct the HEC-HMS rainfall-runoff model; through the screening of different runoff methods, different runoff schemes were established to explore the optimal runoff algorithm for the arid basin; there are mainly SCS-CN method, initial and constant method, Green and Ampt method and exponential loss method in runoff generation; SCS unit hydrograph and Clark unit hydrograph methods are selected for slope transform, kinematic wave and lag algorithms for channel routing, totally 16 different runoff schemes. Three floods in the Gulang River Basin are selected to analyze the simulation effect of different runoff schemes in Gulang River Basin. [Results]1) If the Nash-Sutcliffe efficiency coefficient is selected as the evaluation index, the best simulation results are schemes 1 and 16 in the 20180826 flood, schemes 2 and 1 in the 20190626 flood and schemes 1 and 3 in the 20190911 flood. If the percentage of flood peak error is selected as the evaluation index, the best simulation results are schemes 16 and 1 in the 20180826 flood, schemes 9 and 6 in the 20190626 flood and schemes 13 and 4 in the 20190911 flood.If the percentage of runoff depth error is selected as indicator, the best simulation results are schemes 1 and 8 in the 20180826 flood, schemes 7 and 8 in the 20190626 flood and schemes 6 and 13 in the 20190911 flood. 2)The mean value of Nash-Sutcliffe efficiency coefficient obtained by the SCS-CN method for runoff generation, the SCS unit hydrograph for slope transform and the lag algorithm for channel routing are 0.8, 0.65 and 0.65, respectively; the mean absolute percentage errors of flood peak are 9.29%, 9.71% and 8.47%, respectively; the mean absolute percentage errors of runoff depth are 6.07%, 7.17% and 7.74%, respectively; the mean time difference of flood peak of SCS unit hydrograph for slope transform and lag algorithm for channel routing are 1.21 hour and 1.5 hour, respectively.[Conclusion]The most suitable scheme is the combination of the SCS-CN method for runoff generation, SCS unit hydrograph for slope transform, and lag algorithm for channel routing. The results can provide a certain reference for the prevention of flood disasters in the arid region of Gansu Province.